English
Related papers

Related papers: Some results for poisoning in a catalytic model

200 papers

The recent success of machine learning (ML) has been fueled by the increasing availability of computing power and large amounts of data in many different applications. However, the trustworthiness of the resulting models can be compromised…

Cryptography and Security · Computer Science 2024-03-11 Antonio Emanuele Cinà , Kathrin Grosse , Ambra Demontis , Battista Biggio , Fabio Roli , Marcello Pelillo

It is hypothesized that, within Bose-Einstein condensates, contaminants will form a potential that effects the energy state of a condensate. While assuming a system governed by the Gross-Pitaevskii equation, contaminants are modelled as…

Other Condensed Matter · Physics 2007-05-23 Alexander Oshmyansky

We investigate Turing instability and pattern formation in two-dimensional domains for two reaction-diffusion models, obtained as diffusive limits of kinetic equations for mixtures of monatomic and polyatomic gases. The first model is of…

Mathematical Physics · Physics 2026-02-23 Stefano Boccelli , Giorgio Martalò , Romina Travaglini

Data poisoning causes misclassification of test time target examples by injecting maliciously crafted samples in the training data. Existing defenses are often effective only against a specific type of targeted attack, significantly degrade…

Machine Learning · Computer Science 2022-10-19 Yu Yang , Tian Yu Liu , Baharan Mirzasoleiman

Disorder is everywhere in nature and it has a fundamental impact on the behavior of many quantum systems. The presence of a small amount of disorder, in fact, can dramatically change the coherence and transport properties of a system.…

Disordered Systems and Neural Networks · Physics 2022-06-23 Chiara D'Errico , Marco G. Tarallo

By example of a particle interacting with ideal gas, it is shown that statistics of collisions in statistical mechanics at any degree of the gas rarefaction qualitatively differs from that conjugated with Boltzmann's hypothetical molecular…

Statistical Mechanics · Physics 2015-12-16 Yu. E. Kuzovlev

Data poisoning and backdoor attacks manipulate training data in order to cause models to fail during inference. A recent survey of industry practitioners found that data poisoning is the number one concern among threats ranging from model…

Machine Learning · Computer Science 2021-06-18 Avi Schwarzschild , Micah Goldblum , Arjun Gupta , John P Dickerson , Tom Goldstein

Introduction: Methods now exist to detect residual confounding. One requires an "indicator" with two key properties: conditional independence of the outcome (given exposure and measured covariates) absent confounding and other model…

Methodology · Statistics 2015-10-26 W. Dana Flanders , Matthew J. Strickland , Mitchel Klein

The success of machine learning is fueled by the increasing availability of computing power and large training datasets. The training data is used to learn new models or update existing ones, assuming that it is sufficiently representative…

Preference learning is a central component for aligning current LLMs, but this process can be vulnerable to data poisoning attacks. To address this concern, we introduce PoisonBench, a benchmark for evaluating large language models'…

Cryptography and Security · Computer Science 2025-06-09 Tingchen Fu , Mrinank Sharma , Philip Torr , Shay B. Cohen , David Krueger , Fazl Barez

Regression models are widely used in industrial processes, engineering, and in natural and physical sciences, yet their robustness to poisoning has received less attention. When it has, studies often assume unrealistic threat models and are…

Machine Learning · Computer Science 2026-03-03 Javier Carnerero-Cano , Luis Muñoz-González , Phillippa Spencer , Emil C. Lupu

We discuss the transition from a fully decoherent to a (quasi-)condensate regime in a harmonically trapped weakly interacting 1D Bose gas. By using analytic approaches and verifying them against exact numerical solutions, we find a…

Atomic Physics · Physics 2013-05-29 Isabelle Bouchoule , Karen V. Kheruntsyan , Gora Shlyapnikov

We study the phenomenon of real space condensation in the steady state of a class of mass transport models where the steady state factorises. The grand canonical ensemble may be used to derive the criterion for the occurrence of a…

Statistical Mechanics · Physics 2007-05-23 M. R. Evans , Satya N. Majumdar , R. K. P. Zia

Imperceptible poisoning attacks on entire datasets have recently been touted as methods for protecting data privacy. However, among a number of defenses preventing the practical use of these techniques, early-stopping stands out as a…

Machine Learning · Computer Science 2022-04-20 Pedro Sandoval-Segura , Vasu Singla , Liam Fowl , Jonas Geiping , Micah Goldblum , David Jacobs , Tom Goldstein

A geometrical model which captures the main ingredients governing atom-diatom collinear chemical reactions is proposed. This model is neither near-integrable nor hyperbolic, yet it is amenable to analysis using a combination of the recently…

Chaotic Dynamics · Physics 2018-04-10 L. Lerman , V. Rom-Kedar

By constructing the super-particle representation of the free boson gas, we propose a new statistics in which the particles are non-exclusive. This statistics can be considered as a generalization of Bose-Einstein's. The possible…

Statistical Mechanics · Physics 2007-05-23 Yupeng Wang

Our understanding of various states of matter usually relies on the assumption of thermodynamic equilibrium. However, the transitions between different phases of matter can be strongly affected by non-equilibrium phenomena. Here we…

Quantum Gases · Physics 2013-05-03 Alexander L. Gaunt , Richard J. Fletcher , Robert P. Smith , Zoran Hadzibabic

Low-temperature plasmas are partially ionized gases, where ions and neutrals coexist in a highly reactive environment. This creates a rich chemistry, which is often difficult to understand in its full complexity. In this work, we develop a…

Plasma Physics · Physics 2024-09-13 Diogo R. Ferreira , Alexandre Lança , Luís Lemos Alves

We present a very simple model of a spontaneous emission from a two-level atom, interacting with a field of a finite number of states. Such a process is often said to occur because of the large number of equally-probable states of…

Quantum Physics · Physics 2012-04-03 Krzysztof Piotr Wójcik

It has been shown recently that Bose Gase with weak pair (enough well) interaction is non ergodic system. But Bose Gase with weak pair interaction is so general system that it is evident that the majority of statistical mechanics systems…

Statistical Mechanics · Physics 2011-10-18 D. V. Prokhorenko